On Always being Right (about What One is Thinking)

I Introduction

There are a number of strands to the knowledge we have of our own minds; two strands are these: we often know with ease what we are thinking and we often know with ease what it is we believe. This paper concerns the knowledge of what we are thinking; it pursues questions as to what kind of judgment subjects make about their own thoughts, how those judgments are formed and why they constitute knowledge; it also asks how these judgments relate to the judgments subjects make about their own beliefs when they know with ease what they believe. It focuses on the account developed by Tyler Burge (1988, 1996, 2003) as part of his project of reconciling externalism about thought content with privileged self-knowledge. Burge’s account is well known and influential; as such it is a fitting target for examination and criticism.

By choosing these questions as those to pursue, I am already making some assumptions about self-knowledge — I am assuming that knowledge of our own thoughts is continuous with knowledge of other things, in the following sense. When investigating the nature of subjects’ knowledge of X, one asks about what judgments subjects make about X, about the truth-conditions of those judgments, about the faculty by which these judgments are made, and about the reliability and so on of that faculty. The assumption I am making about self-knowledge is that it is to be investigated in just the same way — one knows what one’s thoughts are by making knowledgeable judgments about one’s thoughts, and so the way to investigate this kind of self-knowledge is to investigate the nature of these judgments.

Burge calls the judgments that constitute subjects’ knowledge of their own thoughts cogito-like judgments, and he defends the thesis that cogito-like judgments are self-verifying — let’s call that thesis Burge’s Thesis.¹ Burge’s Thesis will be at the centre of my investigation of the nature of the knowledgeable judgments subjects make about their own thoughts. I will ask whether there is any account of the nature of these judgments — about their truth-conditions, and about the nature of the process by which these judgments are made — on which Burge’s Thesis is true. Most of this paper will be spent investigating whether Burge’s own views yield an adequate account. I will conclude that Burge’s own view is crucially unclear; he offers hints at a direction in which a careful account might be developed, but I argue that this account would be fatally flawed. I end by sketching an account that looks like a far more promising way to develop Burge’s suggestions.

II Cogito-like judgments

Cogito-like judgments are a sub-class of propositional attitude self-ascriptions. A propositional attitude self-ascription is a judgment that ascribes a propositional attitude to oneself. Propositional attitudes are typed along two dimensions: attitude and content. So a belief that the world moves and a belief that water is a liquid are tokens of the same attitude-type — the type belief — but different content-types. In contrast, a belief that it is not raining and a hope that it is not raining differ along the attitude dimension but are two tokens of the same content-type (or, as I shall say as shorthand for this: they share the same content — the content it is raining ).² When I talk of a self-ascription of a propositional attitude, I mean a judgment which ascribes a propositional attitude to oneself, typing along both the attitude and the content dimensions. So the judgment that I believe that water is a liquid is an example of a self-ascription; on the other hand the judgment that I am thinking is not a propositional attitude ascription as I mean it here, as it does not specify the ascribed thought along both the attitude and content dimensions (it specifies only the attitude dimension — thinking).

Cogito-like judgments are (present-tense) propositional attitude self-ascriptions that along the attitude-dimension ascribe merely the broad...